ASEAN enterprises moving from ‘AI-first’ to ‘AI-native’

Organisations across Southeast Asia are moving beyond the experimentation phase of artificial intelligence (AI) adoption and are actively reshaping their operations to be AI-native, according to Jeff Johnson, managing director for ASEAN at Amazon Web Services (AWS).

Speaking at a regional media briefing on the sidelines of AWS re:Invent 2025 in Las Vegas, Johnson cited new research conducted with Strand Partners, which found that AI adoption in the region has grown by 38% year over year (YoY).

He noted that 29% of businesses across ASEAN have now adopted AI in 2024, up from 21% a year ago. “That translates to 21 million businesses across ASEAN that are using AI, and that works out to about 11 businesses on average per minute that are beginning to adopt AI technologies,” he said.

While most of these businesses (73%) remain at a basic level of adoption focused on productivity gains, Johnson noted that companies in the region are no longer just bolting AI onto existing workflows. “[Instead,] organisations are moving from the era of experimentation in 2023 to becoming more than just AI-first but AI-native,” he said. “They are really reimagining processes and thinking more from a blank sheet of paper.”

To support the growing demand for cloud and AI services, AWS has been expanding its footprint across Southeast Asia with cloud regions in Thailand, Malaysia and Indonesia, alongside its existing Singapore hub. Johnson pointed out the scale of the company’s commitments, noting a projected $9bn investment in Singapore through 2028 and about $5bn in Thailand.

The region’s potential, however, isn’t measured on hyperscaler investments alone, but also in terms of demographics and talent. With 61% of ASEAN’s population under the age of 35, Johnson described the tech-savvy, mobile-first region as a “crucible” for digital innovation.

He also noted that more overseas-educated talents are returning to the region to build startups, challenging the dominance of traditional Western tech hubs. “We’re seeing what I would describe as almost a very positive reverse brain drain," he said.

At the same time, the region is seeing the rise of indigenous and sovereign large language models (LLMs). Johnson pointed to the Typhoon model, an open source Thai-language LLM based on Mistral that was built by SCB 10X, the investment arm of the SCBX financial technology group, as an example of open-source technology meeting local needs.

"We’re seeing a real proliferation of models that build on top of open-source models," Johnson said, noting that organisations are doing so to support the different languages and cultures across ASEAN.

Against this backdrop, AWS is looking to democratise AI development through tools such as Kiro that lets non-engineers build applications with the help of coding agents. At Singapore Polytechnic, for example, students with no coding experience can use Kiro to create products and get involved in real-world projects with AWS customers and partners as part of the AWS AI Spring programme.

Johnson also pointed to several regional customers who are already shifting to AI-native operations. In Malaysia, stock media platform 123RF is using Amazon’s Nova Pro and Nova Lite models to tag over 230 million images and videos, which has reduced customer search time by 90%.

In Thailand, real estate developer Sansiri is processing more than 50,000 invoices a month with 90% accuracy using generative AI, effectively halving their document processing time. And in the Philippines, Union Bank is using Amazon Quick Suite, a generative AI-powered business intelligence platform, to democratise data access for over 200 business users, which Johnson said has sped up access to insights by three to five times.

Despite the optimism, Johnson acknowledged that skills remain a primary barrier, with 52% of businesses citing it as a hurdle to AI adoption. Still, he remained bullish about the region’s potential, concluding that ASEAN is at the forefront of the AI era as organisations use available tools to leapfrog legacy constraints.